Our laboratory has a long-standing interest in non-Pgp mediated mechanisms of drug resistance, having established several cell line models of resistance focusing on the ABC half-transporter ABCG2. We successfully cloned ABCG2 from a mitoxantrone-resistant colon cancer cell line, S1-M1-80, that exhibited an ATP-dependent reduction in drug accumulation. Comprising 6 transmembrane domains and a single ATP binding domain, the gene encodes a half-transporter molecule, and it is thought that dimerization is required for activity. Overexpression of ABCG2 renders cells resistant to mitoxantrone and to the camptothecins, topotecan and SN-38 (the active metabolite of irinotecan). Both substrates and inhibitors of ABCG2 have been discovered at an accelerating pace, and the variety of both substrates and inhibitors rivals that described for P-glycoprotein. There is increasing evidence supporting a role for ABCG2 in limiting the oral absorption of pharmacologic agents and in limiting the brain uptake through expression in the blood brain barrier. We have worked on structure and function relationships in the protein. We identified a drug-induced mutation in ABCG2 (R482T;R482G) that alters substrate and inhibitor specificity;and then carried out a sequence analysis of ABCG2, identifying single nucleotide polymorphisms. We and others reported impaired transport in cells bearing a single nucleotide polymorphism at amino acid 141 that changes glutamine to lysine. Since gastrointestinal absorption of topotecan has been related to ABCG2 expression in the intestinal epithelium, one implication of this work is that the Q141K SNP could be associated with increased exposure to substrate drugs in patients. To the extent that ABCG2 is involved in drug exposure, this SNP could increase exposure to substrates such as imatinib, irinotecan or topotecan. Further, ABCG2 is expressed in the endothelial cells in the brain and another important role for the protein is in protection of the CNS as a component of the blood-brain barrier. An implication of this finding is that compounds that circumvent ABCG2 and thus, the blood-brain barrier, could have increased efficacy in treating or preventing CNS metastases. The importance of localization of ABCG2 in the blood brain barrier will increase as we recognize ABCG2 substrates in new agents such as small molecule tyrosine kinase inhibitors. To evaluate dimerization of ABCG2, our laboratory studied a GXXXG dimerization motif in transmembrane helix 1. We found this motif critical for normal transport activity in ABCG2, but could not prove that it mediated dimerization. We recently discovered that mutation of a nearby residue, T402, completely destabilized the protein. Mutation of the highly conserved amino acid residue 553 (the homologous residue in Drosophila is said to drive dimerization) results in loss of protein expression on the mammalian cell surface, and expression of a nonfunctional protein on the insect cell surface. In both of these systems, chemical cross-linking is preserved, suggesting a proximity of the two monomers even when the protein fails to fold properly. We also identified a critical residue in amino acid 383 that may promote communication between the major domains of the protein. We recently studied mechanisms of regulation of ABCG2, discovering that the ABCG2 promoter is methylated in some renal cell cancer cell lines, resulting in reduced expression of the gene. We also determined that the promoter is regulated by histone acetylation and that depsipeptide is able to upregulate expression in some cell types. Interestingly, there are also cell types where the gene is repressed in a non-HDAC, non-methylation dependent manner. This study may help us understand how normal stem cells turn off ABCG2 as they differentiate and how some cancer cells re-express the transporter;eventually leading to strategies to target ABCG2-expressing cells. We have also learned using chromatin immunoprecipitation that permissive epigenetic marks are evident in the ABCG2 promoter in cells that respond to HDAC inhibition with upregulation of ABCG2 mRNA. However, repressive epigenetic marks persist in the ABCG2 promoter in cells that do not respond to HDAC inhibition with ABCG2 mRNA upregulation. These repressive marks persist in the promoter despite acetylation of the lysine tails of nearby histone proteins, and despite upregulation of other genes in the same cells. This constitutes a model for studying resistance to HDAC inhibitors, an important facet of project #2: II. Clinical and Laboratory Studies of the Histone Deacetylase Inhibitor Depsipeptide. We recently recognized that cells expressing high levels of ABCG2 have truncation at the 3'untranslated region of the gene, which removes a microRNA binding site. The microRNA, hsa-miR 519c, was shown to bind at this site and reduce gene expression and protein translation. New studies have shown that a variety of drug resistant cell lines use the short 3'UTR, suggesting it as a general mechanism of ABCG2 upregulation. Our next aim is to evaluate other micro RNAs involved in drug resistance. In order to evaluate clinical samples for ABCG2 expression, we have developed, in collaboration with Stephen Hewitt and Patty Fetsch, an immunohistochemical assay that has been used to evaluate slides bearing tissue arrays. Using these slides we hope to discover specific tumor types overexpressing ABCG2. In collaboration with Dr. Michael Dean and the Molecular Targets Development Program, led by Dr. James McMahon, ABCG2-overexpressing cells have been used to screen for inhibitors of ABCG2. This is an important undertaking since the potential ability to modulate oral drug absorption and CNS uptake will be important whether or not ABCG2 proves important in oncologic drug resistance. Dr. Curtis Henrich, of the Molecular Targets Development Program, has identified a number of hits that we have now confirmed as ABCG2 inhibitors. These compounds have been evaluated in secondary screens in our laboratory and in that of Dr. Suresh Ambudkar, to prioritize for further preclinical development. In another strategy to identify ABCG2 substrates and inhibitors, we characterized ABCG2 expression in the 60 cell lines of the NCI drug screen. This profile allowed us to identify potential substrates and inhibitors using the COMPARE analysis. These compounds have also been secondarily screened. We have concluded that the drug screen cannot be used to specifically define whether or not a compound is an ABCG2 substrate -- however, a set of compounds can be enriched for substrates by this method. Our plan is to take these compounds into animal studies in collaboration with Dr. William Figg's group. In particular, we hope to assess whether any of these inhibitors will increase CNS accumulation of ABCG2 and Pgp drug substrates. Clinical studies may also be possible using radiolabeled imaging agents (Dr. Pete Choyke and Dr. Jim Doroshow) in combination with inhibitors already available for clinical testing. Finally, we have worked to identify strategies to improve the function of a polymorphic ABCG2, a variant bearing a glutamine to lysine transition at amino acid position 141. Based on our work with the dimerization motif we have studied whether we could mediate pharmacologic rescue of ABCG2 through the addition of substrates. Interestingly we found that the best mediator of rescue was romidepsin. Work is ongoing to understand the mechanism of this rescue. Collectively, this work is best characterized as an effort to both understand and exploit ABCG2 as a therapeutic target to improve anticancer therapy.